Like many elements of our investment industry, market-cap weighted benchmarks have the beauty of being long-standing, familiar and simple. Performance history is available for analysis, we all speak the same language and the indices are self-adjusting as markets appreciate and depreciate. However there are other ways of constructing benchmarks that are elegant and attractive. More importantly, non market-cap weighted benchmarks may offer higher returns and lower risk.
The perils of market-cap weighted indices have been known for some time. In the early 1990’s, as Japan’s market weight dominated EAFE benchmarks, many managers created superior long-lasting track records by the simple expedient of underweighting Japan. In the late 1990’s, Nortel’s growing dominance in the TSE index became a key factor in discussions with Canadian managers, which ultimately led to the creation of the 10% capped index as a means of protecting investors from what was correctly perceived as an accident waiting to happen.
Market cap weighted benchmarks often take on these structural biases with heavy concentration risk. As the market value of stocks appreciate, their share in the market index also increases, which gives rise to fairly large implicit bets. Historically these bets have often imploded with painful consequences.
A “poor man’s alternative” to the market-weighted approach is the equal-weighted portfolio.
A rebalancing strategy needs to be defined to minimize transaction costs since the equal weighted portfolio will develop tracking error over time compared to the market portfolio. However an equal-weighted portfolio approach has the benefits of the following:
- broader diversification and less risk concentration;
- potential takeover premiums (in the case of M&A targets); and
- a fuller replication of sector constituents (when a single large player dominates the market-weighted index).
In my experience, an equal weighted approach can give rise to significant outperformance over market-weighted approaches with very little additional cost or fees. For the more adventurous investor, a number of more sophisticated ways to construct non-market-weighted benchmarks are available.
Rob Arnott is a pioneer in this area, having developed a fundamental index using a process that weights securities based on their economic footprint. Factors include sales, revenues, book values and earnings or number of employees rather than share capitalization. Some of the drivers cited as leading to the success of this approach include a value bias and a rigorous rebalancing policy.
Another approach has been developed by Tobam Asset Management, which develops an index portfolio with diversified risk factors across all sources of risk. This approach avoids index biases towards low or high volatility stocks so as to deliver a more stable performance across market cycles. Investors capture the equity risk premium without exposure to concentration risks, which are fundamentally unrewarded risks.
The arguments in favour of using a new approach to benchmarking are even more compelling when one looks at the fixed income component of the portfolio. When entering into another credit downturn, why would an investor wish to overweight the very companies that have taken on the most debt on their balance sheets?
Whatever one’s views on index construction, it seems only prudent to pay close attention to the composition of market benchmarks, consider a more dynamic approach to creating passive portfolios and ensure that one’s active managers are given a true performance challenge.
Good managers will welcome the chance to prove that their “alpha” isn’t really “lazy beta” and plan sponsors may be able to score a few extra points of return!
